Special Session on 3D Perception and Applications

Session Chair: Homayoun Najjaran, The University of British Columbia, Canada

Abstract: 3D computer vision and 3D image understanding refer to the analysis and recognition of the objects using volumetric images and point clouds. Benefiting from both visual and geometrical information, 3D or shape-based computer vision can outclass 2D or image-based methods in many applications including industrial automation and robotics, autonomous vehicles, and medical imaging. The superiority of 3D computer vision may, in general, be attributed to the richer contents and features of volumetric images and that they are less affected by camouflage, disguise, lighting conditions, image quality, and noise. However, three-dimensional analyses of volumetric images involve a higher dimension and hence prone to error, if not treated properly. With the introduction of impressively accurate and computationally efficient techniques empowered by machine learning, 3D computer vision is increasingly used by researchers in broad applications including industrial automation and robotics, autonomous vehicles, medical imaging, to name a few. This special session invites original work on the algorithmic and applied research on 3D computer vision. Topics may include, but are not limited to:

  1. object recognition, detection, and tracking for unmanned systems applications including autonomous navigation of UAVs and driverless cars,
  2. industrial photogrammetry and metrology that is of great importance to precision manufacturing and industrial automation, and
  3. 3D biomedical imaging using CT Scans and 3D ultrasound images for faster and more reliable screening.